JDK-8268822 : Performance of bulk copy with segments seems suboptimal
  • Type: Bug
  • Component: core-libs
  • Affected Version: 16,17
  • Priority: P3
  • Status: Closed
  • Resolution: Cannot Reproduce
  • Submitted: 2021-06-15
  • Updated: 2022-02-03
  • Resolved: 2022-02-03
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 18
18Resolved
Related Reports
Relates :  
JDK-8269119 - C2: Avoid redundant memory barriers in Unsafe.copyMemory0 intrinsic
Relates :  
JDK-8268743 - Require a better way for copying data between MemorySegments and on-heap arrays
Description
This branch:

https://github.com/openjdk/panama-foreign/compare/foreign-jextract...mcimadamore:small_copy_benchmark?expand=1

Contains a new benchmark "TestSmallCopy". No matter how the copy is done (slicing, or static method), the performance of the segment version is worse than when using unsafe directly.

This could be a result of the added safety check which are added by the segment version, but I believe we should investigate this deeper, as something related seems to be causing issues with Lucene.
Comments
This was addressed by JDK-8269119
03-02-2022